The purpose of this paper is to investigate the determinants of dollar debt and bank borrowing using data from a unique Lebanese firm(level survey that we designed and administered. We also gain novel insight into the financial constraints that hinder the development of companies in Lebanon, and whether and how the political shock in 2005 affected firmsqbusiness conditions. We also compare the robustness of the findings from our data set to those obtained when using the World Bank Enterprise Survey data on Lebanon that was also carried out in a similar time period.
Emerging market crises over the past decade have focused attention on the extensive and possibly excessive use of dollar debt. The resulting currency mismatches have been singled out as an important factor amplifying crises. For example, Aguiar (2005) finds that the investment of exporting firms in Mexico was adversely affected by the peso devaluation in 1994. Firms with high levels of foreign currency debt subsequently decreased their investment compared with other firms. Nonetheless, it is not unambiguous because firms that are seemingly most exposed to a depreciation may also best placed to deal with this risk. Bleakley and Cowan (2008) find instead that the competitiveness effect of a depreciation overall outweighs the negative balance sheet effect. These papers also have important implications for whether monetary policy in dollarized economies should target a peg or allow the exchange rate to float in the presence of liability dollarization (see for example Elekdag and Tchakarov, 2007 who show that a peg is welfare superior to a float).
Financial dollarization has continued to increase all over the world despite declining worldwide inflation (see De Nicolo et al, 2005). For example, the share of dollar deposits in South America increased from 46% to 56% and in Lebanon from 53% to 69% over the period 1996 to 2001. Over the past few years, significant progress has been made in the literature on different theories rationalizing the use of dollar debt. Concurrently, some papers have empirically investigated the factors affecting a firmqs dollar debt. The limitations of the empirical literature, however, are that it derives almost entirely from Latin America and East Asia. Moreover, the existing studies are typically based on the largest and listed companies and they obtain dollar credit from foreign banks and other lenders, a limitation also emphasized by Tornell and Westermann (2002).
The case of Lebanon contributes significantly to this literature in that the major part of credit is intermediated through domestic banks and it is mostly in dollars. Dollar loans are roughly 84% of commercial bank lending in recent years based on statistics from the Central Bank (Banque du Liban) as can be seen in Figure 1. Lebanon has also not experienced the dramatic financial deregulation and ensuing financial crises that possibly contaminate empirical work done in other world regions. The exchange rate has been pegged to the dollar at 1507.5 lira/$ since 1998. The choice of debt denomination in Lebanon has long been freely determined between the the creditor and the debtor. Additional evidence on the importance of bank intermediation in Lebanon is provided by a comparative study of the Middle East and North Africa (MENA) region by Grais and Kantur (2003) that shows that Lebanon has the highest share of banking assets to GDP (272%) and the lowest stock market capitalization (10%) in the region. Moreover, domestic banks dominate the banking system, accounting for 80% of system assets over the sample and their share of banking system assets has only gone up over time (see Mora, 2009). The findings from this study can serve to derive policy implications and to be able to better answer the important question of whether funds are being channeled efficiently. If banks are channeling (mostly dollar) funds to firms based on their collateral and net worth and not on the investment projectqs return, then aggregate growth will be constrained as well.
The large part of borrowing in emerging markets is intermediated in dollars (see for example Eichengreen and Hausmann, 2005). What are the reasons that firms in emerging markets choose dollar debt? One reason is that much of this is borrowing from foreign lenders and foreigners lend in foreign currency because if they were to lend in domestic currency, they would demand a premium due to transaction costs (see Aghion, Bachetta, and Banerjee, 2000, for a model with this assumption. The reason firms resort to foreign lenders in their model is because domestic consumers are willing to lend only up to a maximum amount tin domestic currency tand therefore the residual comes from foreigners). Eichengreen and Hausmann (2005) claim that foreignersq preference for lending in dollars derives from poriginal sinq. International lenders cannot forgive emerging market borrowers for past monetary sins even after considerable time. There are also theories why domestic banks may want to lend in dollars too, which are more relevant for this study. Calvo (2002) reasons that banks with highly dollarized deposits lend to domestic firms in dollars to match the currency composition of their assets to their liabilities. Regulatory constraints on banks to match their portfolios combined with information asymmetry means that banks will be more likely to lend dollars domestically where they are less informationally disadvantaged. In the dual liquidity model of Caballero and Krishnamurthy (2002, 2003), international lenders also lend dollars only against internationally tradeable assets such as foreign currency revenues. As this limited dollar liquidity will not necessarily be in the hands of those agents that need it to for real investment, it will then be intermediated between domestic agents and the distortions that arise are discussed in more detail below).
It is incomplete, however, to focus only on the supply decision by lenders. Firms in emerging markets are also jointly choosing their dollar debt. Following from poriginal sinqor more generally from the incompleteness of financial markets that lead to a failure of uncovered interest parity, lower interest rates on dollar loans than on domestic currency loans provide firms with an incentive to borrow in dollars. However, there is a trade(off because this dollar debt comes at the cost of a balance sheet currency mismatch. This may increase the risk of firms defaulting in the event of a depreciation. Is the dollar debt that is chosen therefore consistent with an optimal level?
Theories on borrowersqchoice of dollar debt can be divided into four main categories. First, some papers (e.g. Burnside et al, 2001) have emphasized various government policies that guarantee (implicitly or explicitly) dollar debt. Some governments may provide systemic bailout guarantees to avoid defaults on dollar debt. Alternatively, the insurance against exchange rate risk provided by a fixed exchange rate policy encourages firms and banks to incur dollar debt, exacerbating asset(liability currency mismatch, and reducing incentives to hedge the associated exchange rate risk. These government guarantees underpin boom(bust cycles in middle(income countries when combined with asymmetric financing opportunities between the tradeable and non(tradeable sectors, as in Tornell and Westermann (2002).
A second strand of literature (see Caballero and Krishnamurthy, 2002, 2003) focuses instead on the private sector and not on misguided government policy as the reason behind dollar debt. A countryqs limited international liquidity is not necessarily in the hands of the entrepreneurs that need it for real investment. Instead it is with other domestic agents such as banks that lend it to entrepreneurs against the value of their domestic liquidity or net worth. Domestic liquidity can take the form of marketable domestic assets such as real estate. But importantly these domestic assets are only marketable domestically and cannot be pledged to international lenders. Caballero and Krishnamurthy emphasize the interaction between this underdeveloped domestic financial market and an international financing constraint that leads firms to borrow too much in dollars compared to the social plannerqs decision. The idea is that agents undervalue insuring against an exchange rate depreciation and therefore initially borrow too much in dollars. The reason is that in those states of nature when more financing is needed, firms will have a depressed demand for credit because of the limited collateral that they can post. Therefore the available dollars will be undervalued because constrained firms cannot pledge all their investmentsqfuture returns to borrow these dollars.
A third strand (see Jeanne 2000) emphasizes that firms may incur dollar debt as a commitment device tto signal to their creditors that they are rgoodstypes. Jeanne also stresses an alternative model whereby firms choose to borrow in dollars because the incentives to commit a high effort level to the project are stronger when debt is in dollars because if the firm fails to achieve high returns, it is then more likely to be terminated. In either case, dollar debt can, therefore, be obtained at a lower interest rate ex ante.
Fourth, firms may have hedging motives to incur dollar debt. This risk management view is that greater foreign earnings, such as those arising from exports, offer a natural hedge for risk(averse firms with dollar liabilities. Risk averse managers may want to keep profits (after debt payments) stable so this means that firms with exports and foreign currency revenue will also be more likely to incur foreign currency debt. In fact, in some countries like Chile, lending in foreign currency is prohibited to the non(tradable sector (see Gulde et al, 2004). Bleakley and Cowan (2008) find support for this view; they find that firms match the currency denomination of their liabilties with the exchange rate sensitivity of their revenues. As a result, depreciations do not lead to an adverse effect on firms with more dollar debt. Finally, leveraged firms that are more susceptible to financial distress may also want to reduce the likelihood of falling into bankruptcy by taking on less dollar debt.
The rest of this paper is organized as follows. In Section 2, we relate the theories discussed to the empirical findings in the literature. We also touch on existing survey evidence from Lebanon. Section 3 sets out the survey method used in this study and describes the data. Section 4 goes over the results in detail. These are divided into the determinants of dollar debt, the extent of financing obstacles and bank borrowing, robustness checks against an alternative data set, and finally the effect of the February 14, 2005 shock on firmsqbusiness expectations. Lastly, section 5 concludes.
Download
PDF Ebook Finance with a Focus on Dollar Debt: Evidence from a Survey of Lebanese Firms
